Yesterday we discussed DLM Phase 1, Data Capture, and how data must be captured for management as it enters the Enterprise environment.  What happens next is Data Maintenance.

Maintenance

Data is bits of information, to build a whole picture, we often need to combines bits of data.  Even more common, is the need to combine multiple sets of data that comes from varying sources with varying format that must be queried in varying ways.

The activities of making data usable and valuable, integration, cleansing,  enriching, etc.,  happens here.   A key component of Maintenance is governance.

The processes we use to perform Maintenance activities should be able to produce valuable data with attestations for data lineage, data control, and data quality (more on this another day).

Example of Data Maintenance Triggering Event

Fictitious Bank  GlobalMe Savings bought 2 companies, Credit Card Asia Company and Mortgage Europe Company

  1. Business Objective:  Bank GlobalMe wants to present a holistic view of one customer who had accounts in the acquired companies to better serve him and gain insights of what products and services can be offered to which existing clients.  The overarching business objective is to compete and win more customers in the marketplace.
  2. Operational Objective:  Bank GlobalMe Finance office must be able to collect, collate, and  use the data in order to perform day to day operations such as Enterprise Resource Planning and Financial Modeling.
  3. Regulatory Compliance objectives:  do business while cost effectively comply with regulatory requirements.

The aim is to unify 3 data sources into one.

Once we build the design of the master data target, the next core activities are known as ETL – Extract, Transform, and Load:

  • Extracting the data out of source repositories using query languages such .QL, SQL, etc.
  • Transforming the data so that table names are sync, field and header names are normalized, and data formatted correctly.  Essentially this is the part where we transform multiple varying forms of data into a clean, richer common blended set.
    (See featured image).   Once this is done we can now
  • Load transformed data into the Target repository.

End result?  Where once the customer have to log into 3 different accounts, he can now see all his accounts from a single interface.  The business can now view their customer portfolio holistically.  And if we did our job well in the maintenance phase, we should be able to produce regulatory reporting along with the required attestations for data lineage, data control, and data quality.

Leave your thoughts on the comments section.  Or feel free to contact me at datafabricblog@gmail.com 

Happy to send an editable version of the 7 phases of a data lifecycle images,  just send me an email with your request.

About the author:

Juliana Carroll is a problem solver, strategic thinker with 15+ years of Fortune 500 consulting experience delivering measurable results that align with both business competitive requirements and regulatory compliance.

Juliana have delivered exceptional results for organizations including Morgan Stanley, Prudential, Merck, Guardian Life Insurance, Blue Cross Blue Shield of Florida, and Deutsche Bank.

Connect with Juliana Carroll via Linked In.